Hello brunogiovs , JRKO and gannaji,
I have placed the order for 10PCB only and as posted above in #305, all have been booked by various people. So none remain with me.
However, I have attached the gerbers herewith, one of you could order and supply the PCB to other interested guys from easyeda.com.
To others who have booked PCB, pl PM me to know my paypal address.
regards
Prasi
I have placed the order for 10PCB only and as posted above in #305, all have been booked by various people. So none remain with me.
However, I have attached the gerbers herewith, one of you could order and supply the PCB to other interested guys from easyeda.com.
To others who have booked PCB, pl PM me to know my paypal address.
regards
Prasi
Attachments
Last edited:
To obtain the lowest distortion values, for each speaker impedance value and at a specific frequency, P2 should be readjusted (in your circuit).
This means that there will be a certain setting of P2 for each load impedance.
The setting of P2 for Zout = 2 ohms will be different (possibly) than the setting of P2 for Zout = 4 ohms or, even, for other impedances. All this is valid for a single frequency of interest (normally, chosen at 1 KHz).
All this that I comment, is valid and verifiable in my circuit. I have not simulated your circuit to ratify if it is also true, although I think it should also behave like that.
Taking as an example the scheme of post 13, I raise the approximate readjustments for each impedance, as to obtain the lowest distortion values at 1 KHz:
THD = 0.000774058% for S = 45.579% at 8 ohms *
THD = 0.00110155% for S = 62.879% at 4 ohms *
THD = 0.0202407% for S = 56.979% at 2 ohms *
* Conditions: input = 2 V RMS
Last edited:
VIRTUAL EVALUATION OF SCHEMATICS
With the help of Multisim 14 and LabVIEW Tools, I made virtually sing the DLH Super Buffer!!!
.
An audio file in .wav format has been injected as an input signal, in the schematic of Multisim 14. Another file in the .wav format was obtained from the same buffer output. Both files are sampled at 44.1 KHz at a resolution of 16 bits, in monophonic.
The files correspond to the input signal, the output signal in single ended mode (trimpot S set to 0%) and the output signal in push pull mode (trimpot S set to 52.8%). Thus, the three signals can be compared.
Although ideal power sources have been used, nothing prevents simulating the possible real ripple of the source that we finally use. This would allow us to evaluate virtually the signal-to-noise ratio.
The same could be applied for the evaluation of crosstalk: that is, assembling two schematics and considering all the parameters of the layout of the specific power supply that we are going to assemble in reality.
The possibilities of evaluation are very interesting, without the need of having any real component. The limit is in our imagination.
To evaluate the possible differences, you can play the files one after the other, with the use of any audio application.
Although very subtle, the differences are obvious.
regards
With the help of Multisim 14 and LabVIEW Tools, I made virtually sing the DLH Super Buffer!!!
.An audio file in .wav format has been injected as an input signal, in the schematic of Multisim 14. Another file in the .wav format was obtained from the same buffer output. Both files are sampled at 44.1 KHz at a resolution of 16 bits, in monophonic.
The files correspond to the input signal, the output signal in single ended mode (trimpot S set to 0%) and the output signal in push pull mode (trimpot S set to 52.8%). Thus, the three signals can be compared.
Although ideal power sources have been used, nothing prevents simulating the possible real ripple of the source that we finally use. This would allow us to evaluate virtually the signal-to-noise ratio.
The same could be applied for the evaluation of crosstalk: that is, assembling two schematics and considering all the parameters of the layout of the specific power supply that we are going to assemble in reality.
The possibilities of evaluation are very interesting, without the need of having any real component. The limit is in our imagination.
To evaluate the possible differences, you can play the files one after the other, with the use of any audio application.
Although very subtle, the differences are obvious.
regards
Attachments
Last edited:
Very interesting DiegoMJ. I will listen to them as soon as I get in front of my headphone rig. It’s in a way similar to what I did in this thread to allow people to hear differences between amps by recording same track with different amps driving a resistive load and recording the signal on that load resistor.
Blind Virtual Audition of Several Headamps
Blind Virtual Audition of Several Headamps
Your work is truly amazing and is far from what I can do!
The technique that I presented here serves to evaluate with acceptable approximation schemes of what we are going to build.
I believe that both techniques can complement each other, if they are circumscribed with adequate methodology.
The technique that I presented here serves to evaluate with acceptable approximation schemes of what we are going to build.
I believe that both techniques can complement each other, if they are circumscribed with adequate methodology.
Your MOSFET models probably don't include subthreshold conduction, so the results may not be sonically applicable. With LTSpice you can use these models:
http://www.diyaudio.com/forums/software-tools/266655-power-mosfet-models-ltspice-12-print.html
Here is a simulation that should create output.wav (which is included in the zip file so you don't have to run the simulation yourself). You will have to set the bias points yourself and run the simulation again if there is something wrong with my choices. I have to use IRFP240 since there is no subthreshold conduction model for IRF150 that I know of.
DLH-sound.zip - Google Drive
I resampled the input file to 176400Hz because the aliasing that would result from the simulator would be much more severe than a rate conversion by a reasonable software resampler. At least now there is the potential in theory for a valid comparison.
http://www.diyaudio.com/forums/software-tools/266655-power-mosfet-models-ltspice-12-print.html
Here is a simulation that should create output.wav (which is included in the zip file so you don't have to run the simulation yourself). You will have to set the bias points yourself and run the simulation again if there is something wrong with my choices. I have to use IRFP240 since there is no subthreshold conduction model for IRF150 that I know of.
DLH-sound.zip - Google Drive
I resampled the input file to 176400Hz because the aliasing that would result from the simulator would be much more severe than a rate conversion by a reasonable software resampler. At least now there is the potential in theory for a valid comparison.
Thanks for the kind words.
I listened to the tracks with my low distortion Aksa Lender preamp driving a very neutral/low THD commercial headphone amp with DT880-250 headphones.
I listened with a self-imposed blind test by mixing the order of the output files until I did not know which was which. The put the tracks side by side in Audicity on loop play and clicked mute/unmute to change tracks while they both played.
I definitely liked the sound of one over the other. It sounded more natural, full-bodied and engaging. No surprise to me when it was revealed to me that I prefer SE sound over PP.
Thanks for the comparison. Neat idea to do it via simulation imposed transfer function.
So this means we should be able to have a digital sound processor imbue any music with a SE Class A amp sound then.
However, you cannot fake a zero global feedback amp this way. They just have a different presentation that cannot be "faked" or simulated. The phase information on ZGF amps is more accurate because there is no feedback to mess it up.
PCB received. All of them will be sent out. PCBs for you and pen for me
.There is a small error on the PCB, the name designations of components are missing
. However, the stuffing guide, in the above quoted post # 309, should help. Its a very unfortunate and silly error on my side, really sorry.If anyone has reservations on this error, please let me know by tomorrow afternoon, I shall refund the cost.
I plan to ship the PCB by tomorrow afternoon.
I have attached revised gerbers (with name designations added), if any one plans to order.
regards
Prasi
